Last Updated 23/08/2019
Ever wondered why, for exactly the same design problem, there’s always so many different solutions? And why no one can agree on what is best? For the same problem, surely we can all follow some logic and science and agree on the same solution?
I remember thinking about this when taking a dump in a toilet in Europe. The toilets in some parts of Europe are designed so that a projected turd hits the ceramic first, preventing splash damage by eliminating high speed impacts with dirty toilet water. That’s good. But then a hot, sticky turd resting on the ceramic will smell more and may get stuck when attempting to flush it down. Bugger, that’s bad. On the other hand Australian toilets yield direct impact with water every time you fire the cannon, resulting in water flying everywhere. American toilets are different again with the swirling vortex created during a flush to help clear all material in the toilet. And of course the Asian toilet is basically a hole in the ground that you squat on. This apparently affords an effortless bowel movement but may not be as comfortable or sophisticated as sitting down.
Why all these options? Why doesn’t the best solution become universal? Even with the differing historical paths that each culture has taken, surely, over time, they should all slowly migrate to the best solution? The answer to these puzzling questions – design compromise. Every design suffers from design compromise. There is never a best solution. There are many ways to skin a cat and any design will always seem wrong in some ways to some people.
In my opinion the idea of design compromise is an important concept to understand. It helps you appreciate that everything in a design is there for a reason. It helps understand that, just because a design is not your cup of tea, it still has merit and may provide advantages for some people in certain applications. It helps you avoid making bad decisions or getting conned by clever marketing people. It helps to avoid wasting money.
Design compromise is a more fancy way of saying:
- You can’t get something for nothing
- If it’s too good to be true it probably is
- There’s no free lunch
- There’s always opportunity cost associated with every decision
No design is perfect. There are always compromises. There are always pros and cons. There are always many adequate solutions. There are always differing opinions on what the most optimal solution is. The design process for a particular item yields just one solution of the many possible solutions.
The Job of an Engineer
I’m an engineer which hopefully gives me some credibility on commenting on what engineers do. It’s how I have come to better understand design compromise. The role of an engineer may be defined as:
Out of the many possible solutions for a design problem, determine which solution provides an adequate solution for the given application.
The parameters of a design sit on a continuous spectrum. There is no black and white. Just shades of grey. The design process involves picking points on the spectrum that suit the application.
Lets design a knife. What material should it be made of? Stainless steel does not rust or stain. But it’s not the best at holding an edge and can chip or break particularly if swung aggressively to chop through something tough. Carbon steel stays sharper and is more resistant to chips and breaks. But it corrodes and stains and always looks dirty. It must not be kept in contact with moisture and must be maintained by cleaning and coating in oil. Which is better? Well it depends on the application. In the kitchen – a nice clean looking stainless steel blade that can be left on a wet chopping board or sink might be better. In the bush? Then maybe carbon steel if the knife is going to be swung around with force. Awesome. We’ve figured out which material is better depending on application. But what if the guy in the bush is working in high rainfall areas or on the beach? Then maybe stainless steel would be better. Bugger, we aint figured out anything! Perhaps we know the kitchen knife should definitely be stainless steel. But what if the guy in the kitchen cuts a lot of super fine sushimi? Or maybe he prefers to buy huge slabs of meat in bulk and cut it all up himself? Cutting through huge slabs of meat with a dull knife is a massive pain in the ass. He might want something that stays sharp, cuts through meat better and is easier to sharpen – carbon steel. How big and heavy should the blade be? Should it be thick or thin? Should it be flexible or rigid? It all depends on the application and user preferences. What is the ultimate knife? There’s no such thing.
If a design is universally better then all other designs become extinct. For example lets consider the design problem of vehicle propulsion and the potential solution of a steam engine. How many vehicles are being produced with steam engines these days? None. Electric and internal combustion engines have universally replaced steam engines in vehicles because they are universally better. Lighter, more fuel efficient, more powerful, more responsive, more compact, more reliable, cleaner. Steam engines have become extinct. Well not quite. There is one application where steam engines are still superior in vehicles – taking tourists on historic train routes. So a few steam engines are still chugging around. Rarely does a design become completely extinct. There’s usually some obscure niche that keeps it alive. But, for a given application, if one design is universally inferior then it will become extinct.
So what can we assume if there simultaneously exists two alternative designs to a particular problem? We can assume that each design has its own advantages and disadvantages. We know this must be true. If one design was universally better, then the other design would not exist. Just through the design existing proves that it must have certain advantages. Even if you do not know or understand what those advantages are, you know that the advantages must exist.
An example related to offroading and four wheel drives where we can apply this idea is the case of leaf springs which I discuss in the article on why I have a Toyota Hilux. Some people believe leaf springs are a stupid outdated design that should not exist on any new vehicle. Understanding the concept of design compromise tells us that is not the case. The fact that leaf springs do exist proves that they must have some advantages over other designs. Without being involved in the design process or having a good understanding of suspension, it’s difficult to know what those advantages are. But the advantages must exist. There must be some compromises between leaf spring designs and other designs. The other designs are not universally better.
A consequence of design compromise is that, to make some feature or performance measure better, you must make something else worse. Nothing is universally better. Even if you do not foresee or understand what you are making worse, if you implement a change to make something better then you invariably make something else worse. Any change could be introducing unintended consequences, particularly if the reasons behind the original design are not known.
Optimising one element of a design effects other elements of the design. You cannot simultaneously optimise all elements of a design. Every part of a design is a compromise so that every feature and performance requirement is simultaneously satisfied. The design process involves understanding how each design element affects each feature or performance measure and then optimising the overall design for the application at hand. No single design element is optimised in isolation.
Tyres are an easy to understand example. Want tyres to last longer? Then you need a harder compound. But harder compounds are less grippy. To make one feature better you must make the other worse. You can’t make a tyre hard and soft at the same time! This means that a tyre will never last as long as it could have and will never have as much grip as it could have. Neither performance measure is optimised. This is true for any design. An individual performance measure is never optimised in isolation and therefore will never be as good as it could be.
So the dude designing the tyre will understand how tyre compound affects grip and wear life, understand the application (is it for road, snow, sand, mud, racing, small car, big car, truck, rich people, poor people, consumer suckers, etc), then come up with the optimal solution for the application (which will always be a compromise between performance measures).
Is it possible to make a superior compound that is both more grippy and lasts longer? Probably, but the tradeoff may be that the superior compound is more expensive. It takes more energy to produce and uses more expensive raw materials. The extra expense may not be worth the improvement in grip and service life. The compromise is cost vs the returns you get for that extra cost. What if there is a compound that lasts longer, is more grippy AND is the same price or cheaper then the old compound? Then the old compound becomes extinct. The design process will eliminate it as an option and there would be no comparison in the first place.
The aerodynamics of a racing car is another easy to understand example. Downforce provides grip and faster cornering but also increases drag and reduces speed on straights. Airflow to the engine also needs to be considered in the aerodynamic design to ensure the engine receives adequate air under all loads and speeds. Driver safety and structural requirements also need to be satisfied. Changing the aerodynamics of a racing vehicle will change downforce, drag, airflow to the engine and potentially interfere with structural and safety components of the car. All these needs must be simultaneously met. The only way to do that is to have a less than optimal design of each isolated design element.
Too Good to be True
If someone is trying to sell you something that offers all advantages and no disadvantages then you should be skeptical. If someone is trying to sell you something that is cheap, simple and amazingly effective then again you should be skeptical. A magnet that saves fuel? A herb that cures cancer? A car that runs off water? A share trading strategy that will immediately make you rich? The concept of design compromise tells us that, if something is cheap, simple and effective, then it should become universally adopted. If there are no compromises involved it must become universally adopted. If something is cheap and simple but not universally adopted then it’s highly likely that it is not effective. The design process automatically adopts any solution that is cheap, simple and effective.
Is the Engineer Always Right?
No. The design process usually yields an adequate solution for the problem at hand. But we’re all individuals and we may prefer certain features over others. Our application may be different to the application envisaged in the original design. There may be some variables that the original design did not account for. There may be some constraints that forced the original design down a certain path. So there is still merit to voicing your opinion on a design and changing stuff. The purpose of this article is not to suggest that engineers are perfect and deserve to be treated like royalty and paid millions of dollars a year (although, for my specific case, I am suggesting that). The point of this article is to explain that there is always compromise involved in any design. There is always compromise involved in any modification to a design. Those compromises should be considered. Be skeptical of anyone offering products that seem too good to be true. Ask yourself why this supposed enhancement is not incorporated in the original design. If it’s cheap, simple and effective why isn’t it universally adopted? If it’s so good why isn’t this guy selling it already a billionaire? Where is his garage full of Ferraris? Avoid blindly spending money on stuff. Invest a bit more time looking into the pros and cons. Consider not spending that money at all so that you can afford to work less and go fishing more.
A lot of this blog is about offroading, camping and fishing so why not list a few relevant design compromise examples?
|Performance chip in your engine||More power||Reduced engine life, higher failure rate|
|Bigger tyres||More ground clearance, tough look, better offroad performance||Reduced fuel economy, poorer on-road performance|
|Suspension Lift||More ground clearance, tough look, better offroad performance||Increased body roll when cornering, increased rollover risk, increased wear rate of CV joints|
|Thermomix||All in one, save time, save space||Ridiculously expensive, sold in a pyramid scheme, complicated motor drive system|
|Solar Panels||Free power, quiet||Initially more expensive per installed watt compared to other power sources, no power when no sun|
|Bigger Exhaust Pipe||More high end power||Less low end power, poorer fuel efficiency|
|Braid fishing line||Stronger, lighter, less stretch, further casting||Higher shock loads, poor abrasion resistance, more likely to tangle and form birds nests|
|Tow a caravan or camper trailer||More comfort, carry more stuff, less overloaded vehicle||More complexity, more maintenance, less enjoying nature, more suppressing your senses, more fuel consumption, inability to access some offroad areas|
|Big fishing rod||Cast further||Awkward to carry, awkward to use around trees and confined areas, tiring to use for long periods|
|Spatulacraft||Performs everything perfectly||No disadvantages, should be universally adopted|